Awesome

NitroFE ( Nitro Feature Engineering )

<a href="https://github.com/psf/black"><img alt="Code style: black" src="https://img.shields.io/badge/code%20style-black-000000.svg"></a>

NitroFE is a Python feature engineering engine which provides a variety of modules designed to internally save past dependent values for providing continuous calculation.

Documentation

https://nitro-ai.github.io/NitroFE/

Installation

Using pip

Use the package manager to install NitroFE.

pip install NitroFE

Available feature domains

Time based Features

Indicator / windows / moving averages features are dependent on past values for calculation, e.g. a rolling window of size 4 is dependent on past 4 values.

While creating such features during training is quite straighforward , taking it to production becomes challenging as it would requires one to externally save past values and implement logic. Creating indicators becomes even more complex as they are dependent on several other differently sized window components.

NitroFE internally handles saving past dependant values, and makes feature creation hassle free. Just use first_fit=True for your initial fit

• Jump right in for a handson

The Time based domain is divided into 'Moving average features', 'Weighted window features' and 'indicator based features'

Indicators based Features

NitroFe provides a rich variety of features which are inspired and translated from market indicators.

• Absolute Price Oscillator
• Percentage price oscillator
• Moving average convergence divergence
• Average true range
• Average directional index
• Aroon Oscillator
• Bollinger Bands
• Kaufman Efficiency
• Triple Exponential Moving Average Oscillator
• Zero lag exponential moving average
• Relative Strength Index
• Inverse Fisher Relative Strength Index
• Series Weighted Average
• Series Weighted Moving Feature
• Keltner channel

Moving average features

In statistics, a moving average (rolling average or running average) is a calculation to analyze data points by creating a series of averages of different subsets of the full data set. NitroFE provides an array of variety of moving averages type for you to utilize.

• Exponential moving average
• Hull Moving Average
• Kaufman's Adaptive Moving Average
• Fractal Adaptive Moving Average
• Triple exponential moving average
• Smoothed Moving Average

Weighted window Features

NitroFe provides easy to use functions to create specified weighted window featuresm and send custom operations as per your need